Jensen Huang: The Doomer Hoax, Superintelligence Is Here, and The Future of AI (ft. President Trump)

Original Media · 47m · English · Podcast · All-In with Chamath, Jason, Sacks & Friedberg

TLDR

NVIDIA CEO Jensen Huang joins the All-In Podcast for a wide-ranging discussion on AI safety, regulation, open vs. closed models, China's AI progress, and NVIDIA's ecosystem strategy. The conversation is interrupted by a live phone call from President Trump, who dismisses AI doom narratives as a 'hoax' and pledges to accelerate US data center and energy infrastructure. Huang argues that AI safety and innovation are not mutually exclusive, that most dramatic AI predictions have been wrong, and that the real 'race' is about who best exploits the technology — not who invents it first.

For you

PRO members can sign in and request more information — or their own version of this digest, tuned to what they care about.

Sign in to unlock

Surprising Insights

•

President Trump called in live, declared AI doom scenarios a 'hoax,' and framed data centers as 'the oil of the next 20-25 years' that make communities wealthy.

•

Jensen Huang describes NVIDIA as 'surprisingly uncompetitive' — they actively avoid competing with customers and prefer enabling the entire ecosystem.

•

China now produces the majority of open source AI contributions by volume due to massive STEM graduate output, but Huang argues downloaded models become 'yours' to fork and improve.

•

NVIDIA builds frontier models in 5 domains (self-driving, biology, etc.) not to disrupt but because customers like Lilly and Mercedes need them and can't build them alone.

•

Huang believes AGI/superintelligence already exists in narrow domains: autonomous driving (1/10th human accident rate) and protein synthesis/virtual screening.

Key takeaways

•

AI safety and rapid innovation are false choices — America can lead and do it safely through engineering controls, not fear-based regulation.

•

Nearly all high-profile AI predictions (radiologist replacement, 90% code generation, mass white-collar job loss) have proven demonstrably wrong, undermining credibility of 'trust the experts' narratives.

•

Open source models are essential for broad-based AI adoption: 80% of $400B in recent AI venture funding went to companies using open models, enabling startups, researchers, and every industry to participate.

•

The AI competition is about exploitation and deployment, not invention — just as the US exploited European electrical inventions to win the last industrial revolution.

•

NVIDIA's strategy is to 'go up as far as we need to and as low as possible' — building foundational infrastructure and frontier models only where customers have unmet needs, then letting 1,000 flowers bloom.

Notable quotes

“A GPU is like a time machine because it lets you see the future sooner.”

▸ Notable quotes · 0:42

“Safety and leadership are not false choices. You're able to innovate quickly. You're able to execute quickly. And America is able to lead and to do it safely.”

▸ Notable quotes · 2:26

“First of all, we shouldn't because it's made up.”

▸ Notable quotes · 4:53

“The race is really about who exploits the technology best.”

▸ Notable quotes · 19:34

“Water is free... You use the right water in the right places. And this is no different than electricity.”

▸ Notable quotes · 16:51

“Our strategy is go up as far as we need to and as low as possible.”

▸ Notable quotes · 37:21

“The robots will not be taken over. The AI will not be taken over the rest of the world. The whole thing is a hoax.”

▸ Notable quotes · 26:25

“Whoever wins, AI wins. That's how big it is. It's bigger than the Internet.”

▸ Notable quotes · 26:56

“We're already there [AGI].”

▸ Notable quotes · 44:56

Chapters

•

0:00 — Introduction and Jensen Huang Welcome: Hosts introduce Jensen Huang as founder/CEO of NVIDIA, calling him 'GPU Jesus' and noting NVIDIA's 97% YoY revenue growth and status as the most important stock in the market.

•

1:30 — Dario Amodei Essay and AI Safety Debate: Discussion of Anthropic CEO Dario Amodei's essay on AI safety, whistleblower concerns, and the conflation of internal control issues with speculative extinction predictions. Huang argues safety and innovation are not false choices.

•

7:00 — Failed AI Predictions and Accountability: Huang lists multiple high-profile AI predictions that proved wrong: radiologist replacement, 90% AI-generated code, 50% entry-level job losses, GPT-2/Llama-3 safety fears. Calls for accountability for fear-mongering.

•

13:00 — AI Regulation and Engineering Controls: Huang argues regulation should solve actual problems, which so far have come only from frontier labs due to their compute advantage. Root-cause engineering fixes — not broad regulation — are the answer.

•

16:00 — Open Source vs Closed Source Models: Huang uses 'bottled water vs tap water' analogy: both closed and open models needed. Open models enable sovereignty, privacy, and 80% of AI startups. China contributes most open source volume but models become forkable property.

•

20:00 — China's AI Progress and the 'Race' Narrative: Huang reframes the AI race as about exploitation, not invention — citing US exploitation of European electrical inventions. China is pragmatic, not doom-focused. Chinese lithography likely viable by 2030.

•

23:00 — President Trump Phone Call: Trump calls in live, calls AI doom a 'hoax,' praises data centers as wealth creators, pledges to stop permitting obstruction, claims $20T investment incoming vs $1T under Biden, and says 'whoever wins, AI wins.'

•

28:30 — NVIDIA Ecosystem Strategy and Capital Allocation: Huang describes NVIDIA as 'bank of AI' — removing bottlenecks across land, power, shell, and supply chain. Strategy: enable entire ecosystem, not compete at application layer. 'Go up as far as needed, as low as possible.'

•

39:00 — Competitive Threats and China Lithography: On Elon's TerraFab: 'If anybody could do it, he can.' On Chinese lithography: volume production strength, likely viable by 2030. Huang thinks in decades — 'two or three years is just a click.'

•

44:00 — AGI, Superintelligence and Closing Thoughts: Huang believes AGI already exists in narrow domains (autonomous driving, protein folding). Superintelligence is next waypoint. Urges toning down drama: 'We're going to make it. We need all of America to come with us.'

Speakers

•

Jensen Huang — Founder, President, CEO of NVIDIA

•

Chamath Palihapitiya — Co-host, All-In Podcast

•

David Sacks — Co-host, All-In Podcast

•

Jason Calacanis — Co-host, All-In Podcast

•

David Friedberg — Co-host, All-In Podcast

•

Donald Trump — President of the United States (phone call)

⋮